Prompt to perform? Sustainable talent management in the age of generative AI

“ChatGPT is taking over tasks that we used to delegate to junior staff.” We currently hear statements like this often in the context of generative artificial intelligence (AI) and work. What sounds like efficiency, however, would have unintended consequences: it is through these tasks that juniors build up the expertise they will later need to use AI effectively. If entry-level tasks were increasingly delegated to AI without considering the impact on talent management, the talent pipeline risks drying up in the long term. This blog post discusses the role that expertise plays in the effective use of AI and how companies can deploy generative AI in a way that strengthens, rather than threatens, their talent pipeline.

Since generative AI tools such as ChatGPT and Copilot have been increasingly used in many office jobs, debates on AI in the workplace have frequently focused on the implications for those starting their careers, such as how entry-level employees are changing the way they work with generative AI or how they deal with the uncertainty of the technology (Kellogg et al., 2025; Mayer et al., 2025). Although assessing which tasks AI is particularly useful for and where it achieves better quality results is complex (Dell’Acqua et al., 2026), employees often perceive it primarily as a support for repetitive, low-complexity tasks; exactly the kind of tasks that, in the past, were frequently done by junior staff. 

It is therefore not surprising that, when speculating about which jobs management could cut thanks to generative AI, the answer is currently often “entry-level jobs”. While AI offers a convenient explanation for losses of entry-level jobs, the underlying causes are often far more complex. For example, in tough economic times, companies tend to hire fewer people rather than lay off existing employees — and so entry-level professionals are the first to feel the impact. 

These and similar kinds of contexts are what we investigate in the research project “Generative AI in the Workplace” (GENKIA). Specifically, we examine how employees experience the use of generative AI. In interviews with professionals from various fields, we repeatedly encounter speculations about how generative AI might change the way people enter the workforce in the future. 

Young talent does not emerge out of the blue

Talent management offers a useful perspective on these developments. It is a subfield of human resource management that focuses on how companies can fill critical positions with suitable employees in the long term. In addition to talent acquisition, this includes, among other things, talent development, performance appraisals, and succession planning. A key part of this is developing young talent. The focus is particularly on those employees who demonstrate above-average performance. These so-called “high performers” or “star employees” are also often more visible, better connected, and more frequently considered for leadership positions (Call et al., 2025). One of the goals of talent management is therefore to identify these individuals early on, provide them with targeted support, and retain them in the company over the long term. But high performers do not emerge out of the blue. They are the result of a functioning talent pipeline that begins with entry into the workforce.

How do young talents build expertise?

New hires initially take on clearly defined, relatively simple tasks, gradually acquiring the specialist knowledge and experience that qualify them for more complex roles. Over time, they grow into more demanding roles, take on responsibility and pass on their knowledge to less experienced colleagues. Companies thereby ensure that critical positions can be filled in the long term, even if experienced employees leave the company or retire.

The key point here is that qualified young professionals with solid practical experience do not just appear at the flick of a switch. They often develop their specialist knowledge and judgement through the supposedly simpler tasks of their first few years in the job. There are different findings on whether and to what extent generative AI could actually be used to displace entry-level jobs – there is however a growing body of evidence suggesting that economic factors, above all, influence the availability of entry-level jobs. Some practitioners in our interviews nonetheless share the view that generative AI primarily supports them in exactly those tasks typically done by juniors. If these tasks disappear without considering how this affects the training of junior staff, there may be a lack of young talent tomorrow.

Expertise is also one of the prerequisites for using generative AI effectively. This raises another question that is relevant to both companies and young professionals: Who actually benefits from generative AI, and under what conditions? 

Employees benefit differently from AI

There are varying findings regarding which employees in companies benefit most from the use of AI. A study on the use of chatbots in customer service, for example, showed that less experienced and less qualified employees were able to improve both the speed and quality of their results using chatbots. The most experienced and highly qualified customer service staff, on the other hand, were only able to increase their speed slightly and actually saw a slight decline in the quality of their results (Brynjolfsson et al., 2025). In a field experiment conducted at a telemarketing firm, on the other hand, it was primarily higher-skilled employees who benefited from being able to use AI support to answer customers’ questions more creatively during sales calls, thereby generating higher sales (Jia et al., 2024). In an online experiment in which participants had to complete a writing task, ChatGPT was able to reduce the disparity between participants (Noy & Zhang, 2023).

Still other researchers argue that, due to the increasing use of generative AI, we will see an “AI-specific Matthew effect” in the future (Call et al., 2025). The Matthew effect states that success leads to more success (“to those who have, more will be given”). In the context of AI, this means that high performers can derive disproportionate benefits from AI, as they can use AI strategically while leveraging their superior expertise. 

What does this mean in concrete terms? Those who already possess comprehensive expertise can use AI more effectively and thereby achieve even better results. Performance gaps may therefore widen, not narrow. 

Experience and specialist knowledge are key for using AI effectively

In our own interviews, practitioners from human resource management emphasize just how important specialist knowledge and practical experience remain. Especially in times of generative AI, employees need a basic specialist knowledge to be able to critically evaluate AI-generated results.

 “This makes it all the more important […] that the person who […] uses this AI […] has the practical experience to verify all the results,” explained an HR consultant to us. 

A specific example: If a recruiter uses AI to draft a job ad and it sets out unrealistic requirements such as 15 years of professional experience for a junior role, the recruiter’s experience allows them to quickly recognize that such requirements are unrealistic and will discourage suitable candidates. Those who lack this experience run the risk of overlooking such errors. To benefit from generative AI in the workplace, employees therefore need not only technical skills but also domain-specific knowledge.

Unintended consequences: Widening performance gaps?

And here lies a consequential irony: Generative AI is increasingly expected to take over exactly those tasks through which entry-level professionals have traditionally built the expertise they later need to use AI effectively. Those without a solid foundation of expertise cannot critically evaluate AI results or use AI strategically. And those who cannot do so fall behind those who already have this knowledge. For the latter – according to the AI-specific Matthew Effect (Call et al., 2025) – benefit even more from AI. Thus, what appears to be an increase in efficiency in the short term undermines the development of specialist knowledge in the long term, and thus the foundation on which future career steps become possible.

Empty talent pipelines as a result of misguided AI strategies

What would be the consequences if, in the future, young professionals found it harder to acquire the necessary specialist knowledge and experience? Or if companies were actually to create fewer jobs for career starters in the hope of carrying out their tasks more cheaply using AI?

While this might save companies money in the short term, in the long run it has far-reaching consequences for both society and businesses. For young people, this means growing uncertainty about which educational and career paths will still be reliable in the future. For companies, it means that their talent pipeline is at risk of drying up. After all, those who do not train career starters today will have no experienced specialists tomorrow. Specialists cannot be produced at the flick of a switch.

A manager in HR development described the changed circumstances for training young talent in one of the interviews for our GENKIA project as follows:

“It is […] a vicious circle, because in every profession […] you always need the basic knowledge before you can […] take on new areas of responsibility. […] Presumably, it will still be graduates [who check AI results in the future], but they will enter the workforce at a different level or simply acquire a different set of foundational knowledge than today’s graduates do.”

The quote shows: It is not just a question of whether entry-level jobs could be eliminated, but also of how companies need to rethink talent management if entry-level jobs undergo fundamental changes.

Outlook: Long-term planning rather than short-term cost savings

The use of generative AI does not necessarily undermine learning processes. It can also promote them. For instance, the aforementioned study on the use of chatbots in customer service showed that AI-based support can not only increase work speed but also generate learning outcomes, for example by improving customer service staff’s English language skills (Brynjolfsson et al., 2025). Generative AI tools can also support talent development through the creation of personalized training and development pathways. 

Companies therefore face the challenge of deploying generative AI in a targeted manner so that it promotes learning and development without creating long-term gaps in specialist knowledge and the talent pipeline. The key question, however, is not whether companies use AI, but how. In practical terms, this means that when AI is used for repetitive entry-level tasks, the time saved should not simply be cut from the budget, but used to involve new recruits in more complex tasks at an earlier stage and to offer them more responsibility. 

At the same time, targeted measures such as mentoring programmes are needed to ensure the transfer of specialist knowledge. After all, if junior staff rely on ChatGPT and similar tools instead of asking experienced colleagues, companies run the risk of misinformation increasingly circulating. Furthermore, companies, much like educational institutions, face the challenge of promoting both the confident use of AI and critical thinking without AI. Incentives for this can be created by considering this in performance appraisals. Ultimately, it is crucial that companies plan for the long term and keep talent management in mind, rather than focusing on short-term cost savings.

References

Brynjolfsson, E., Li, D., & Raymond, L. (2025). Generative AI at work. The Quarterly Journal of Economics, 00(00), 1–54. https://doi.org/10.1093/qje/qjae044

Call, M. L., Jiang, K., & Idso, C. (2025). Star Advantage: Employee Value Creation and Capture in the Age of Artificial Intelligence. Human Resource Management, 65(1), 151–167. https://doi.org/10.1002/hrm.70023

Dell’Acqua, F., McFowland, E., Mollick, E., Lifshitz, H., Kellogg, K. C., Rajendran, S., Krayer, L., Candelon, F., & Lakhani, K. R. (2026). Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of Artificial Intelligence on Knowledge Worker Productivity and Quality. Organization Science, 37(2), 403–423. https://doi.org/10.1287/orsc.2025.21838

Jia, N., Luo, X., Fang, Z., & Liao, C. (2024). When and how artificial intelligence augments employee creativity. Academy of Management Journal, 67(1), 1–60. https://doi.org/10.5465/amj.2022.0426

Kellogg, K. C., Lifshitz, H., Randazzo, S., Mollick, E., Dell’Acqua, F., McFowland, E., Candelon, F., & Lakhani, K. R. (2025). Novice risk work: How juniors coaching seniors on emerging technologies such as generative AI can lead to learning failures. Information and Organization, 35(1), 1–21. https://doi.org/10.1016/j.infoandorg.2025.100559

Mayer, A.-S., Baygi, R. M., & Buwalda, R. (2025). Generation AI: Job Crafting by Entry-Level Professionals in the Age of Generative AI. Business & Information Systems Engineering. https://doi.org/10.1007/s12599-025-00959-x

Noy, S., & Zhang, W. (2023). Experimental evidence on the productivity effects of generative artificial intelligence. Science, 381(6654), 187–192. https://doi.org/10.1126/science.adh2586

Picture: Hanna Barakat & Cambridge Diversity Fund / https://betterimagesofai.org / https://creativecommons.org/licenses/by/4.0/